New visuospatial associations by training verbospatial mappings in the first language

Abstract

An important issue in the study of perception and action is the relationship between the representations of different types of spatial information (e.g., arrows, locations, spatial words). Spatial compatibility effects have been observed for several kinds of spatial information, suggesting similar processing pathways (Fitts & Seeger, 1953). A spatial compatibility effect is the observation that reaction times (RTs) are faster and error percentages lower when participants make a spatially compatible response to a stimulus (left stimulus location requires left response and right stimulus location requires right response), relative to an incompatible mapping condition. Similar effects have been observed when spatial information is task-irrelevant. For instance, when participants press a left or right response key depending on the stimulus color, RTs are faster when the irrelevant stimulus location (left or right) corresponds to the response location. This is typically referred to as the Simon effect (Simon, 1990).Whether the spatial information is activated by arrows, words, locations, or even number magnitude, the perceptual left seems to be associated with the motor left. In this study, we investigated whether different types of spatial information activate a shared spatial representation, instead of directly activating a spatially corresponding response. These shared representations introduce an intermediate or semantic level between the perceptual input and the responses.Intermediate codes are assumed in the computational model of Zhang, Zhang, and Kornblum (1999) in order to explain (spatial) Stroop effects. When two spatial inputs activate the same representation (upward arrow presented above fixation), RTs are fast and error percentages low, as compared with when the two sources activate different representations (e.g., an upward arrow presented below fixation; see, e.g., Liu, Banich, Jacobson, & Tanabe, 2004). In the model of Zhang et al., this effect is explained in terms of possible competition between two intermediate representations. Note that other models for Stroop effects do not implement intermediate codes (e.g., Cohen, Dunbar, & McClelland, 1990).Although the implementation of intermediate shared representations results in an elegant architecture, surprisingly few efforts have been made to test this hypothesis. In this study, we tested the hypothesis of intermediate spatial representations at which different types of spatial information converge. Particularly, we asked whether verbospatial (the words LEFT and RIGHT) and visuospatial information (left and right screen positions) share a common intermediate code.We used a paradigm that was introduced by Marble and Proctor (2000) and Proctor, Marble, and Vu (2000). In this paradigm, participants switch between different tasks as a function of the stimuli. For example, Marble and Proctor instructed participants to respond spatially incompatibly to the location of white stimuli (the inducer task) and to respond left or right on the basis of the color of...

Shared spatial representations for physical locations and location words in bilinguals’ primary language

“…For the first interaction, the Simon effect was larger in the last trial block (M 21 msec) than in the first three blocks (M 5 msec), signifying a reduction in the effect of the location-word mapping similar to that reported by Notebaert et al (2007). The latter interaction reflects that RT decreased more across blocks when the Simon trials were mixed with English words (MD 99 msec) than with Spanish words (MD 43 msec).…”

“…Kroll and Stewart's (1994) model of bilingualism assumes that words from both languages activate a shared conceptual representation, with activation from the first language being stronger than that from the second language. Notebaert et al (2007) proposed a model for processing of spatial information that postulates this conceptual representation's being also activated by visuospatial stimuli. They tested this model by taking advantage of the finding that when trials with incompatible mapping of the words left and right to left and right keypresses are intermixed with trials on which stimuli occur in left and right locations but responses are to be based on stimulus color, the Simon effect for the latter trials is eliminated (Proctor et al, 2000;Proctor & Vu, 2002).…”

“…The design of Experiment 1 followed that described in Notebaert et al (2007) closely, with the main difference being the languages of the bilingual participants and stimuli and the inclusion of a monolingual control group. The bilingual participants were fluent in both Spanish and English, whereas the monolingual participants were fluent only in English.…”

“…This influence of the incompatible location-word mapping on Simon task performance suggests that verbospatial information and visuospatial information share a semantic/conceptual representation, with stimuli in the Simon task tending to activate the response linked to the shared representation for the task of responding to location words. Notebaert et al (2007) showed that, for bilingual participants, this influence of an incompatible location-word mapping on the Simon effect occurred only when the words were from the first language. For bilingual Dutch/ French participants, the Simon effect for the locationirrelevant stimuli reversed in the first of four trial blocks when the words were in the participants' first language For monolinguals, the Simon effect is eliminated when Simon task trials are intermixed with ones in which participants respond to the words left and right with incompatibly mapped keypresses.…”

“…Evidence of processing asymmetry for spatial information in the first and second languages was reported by Notebaert, De Moor, Gevers, and Hartsuiker (2007), using a method in which trials of a Simon task (Simon, 1990) for which stimulus location was irrelevant were mixed with trials of a task for which the meaning of a location word was relevant. For the Simon task, left and right keypress responses were assigned to the colors of red and green circles, presented randomly in left and right locations.…”